THE HISTORY AND SHAPE-OF THE TEST-PIECE. 319
rolled into angles and the charges are grouped according to the thickness of the finished material. The-- great inferiority of the; tests from the 6-ineh ingot is easily explained. It is difficult to cast small ingots so that they will not be scrappy, and the bars rolled from them will oftentimes contain flaws; consequently, we break down the ingot to a billet two inches square and chip out the flaws, after which the piece is reheated and gives a perfect bar. It does not receive sufficient work to ensure. good elongation, but only the strength of the material is under investigation, and in this respect the results are found to be comparable with the finished material.
SEC. XV'Ig.—Comparison of rounds and. flats.—The properties of a flat bar are different from those of a round.
The points involved are three: ..'....'
(•!•) The percentage of work on the piece;
(2) The finishing temperature. ,,, •... '
(3) The shape of the piece.
(1) The amount of reduction from the bloom or ingot should not play too gf eat a part in. the problem, for it is the duty of the manufacturer to so conduct the operation that every piece, no matter how large, shall have sufficient work. But a large section, a 9-inch round,- for example, cannot possibly be finished .under the same thorough and permeatiye compression that, can be put upon a bar only one inch in diameter or upon a thin flat.,
'(2) It is.for the rolling-mill to arrange that every piece is rolled at a proper temperature, but it is impracticable to finish bars of all thicknesses under identically the same conditions.
;(3) The shape of the test-piece has an influence upon the results, but it is difficult to isolate this relation from the' effect of work and finishing temperature. '' . :
'The separation of these three intertwining influences'is a complicated problem, the nature, of which will be illustrated by Table XVI-I, which gives the'results obtained from a large, number of heats by cutting two billets from the same ingot and rolling one. into 'a round and the other into a flat. This table discloses the following facts:
(1) Taking both natural- and annealed bars., there are 18. comparisons between rounds, and flats: The ultimate strength is less Ln: the flat, in, every case." The elastic limit; falls iiT 17 cases, and